A third episode option for accuracy about crunchy-cli

[ohitstom]

I've added a option which should handle the most cases. You can test it by cloning the repo and checking out the feature/relative_sequence_number branch or using one of the pre-built binaries.
The new format option is called {relative_sequence_number} as it is based of the sequence_number api field but it might get renamed when I merge it, to state more clearly what it does.
It would be great if you guys could test this and give feedback @BlastBolt5 @alazare619 @ohitstom

I'll try and find some time to test it sounds awesome tysm!

[BlastBolt5]

Thank you so much, I have been busy with school so I haven't been able to test this yet but it looks perfect

[alazare619]

This is not possible currently.

I'll elaborate on a better solution expanding on your issue. And how to fix this the correct way.

Sonarr/Plex and many other services depend on thexem.info

thexem.info has mappings of anidb to tvdb and to scene/p2pscene releases. This is what sonarr/plex use for metadata mapping.

Notice in that list anidb.

Well anidb also happens to have a public api, said API also has the ability to list a shows and match titles to the title in crunchyroll. This is because anidb has the url of hte japanesse version of each episode in its DB.

If crunchy-cli just allowed us to connect to anidb. It can pass the url of the japanesse audio stream of the current episode your downloading to anidb and get the correct episode/season number of said and return it. This would 100% correct all issues with numbering when doing entire series archiving

[alazare619]

@bytedream

[BlastBolt5]

an option to have relative episode number with out episode 0s or .5s would also be a fix to the main issue, but using that anidb connection would also be amazing.

[bytedream]

A AniDB integration would be indeed nice.
As far as I can see, anidb has a field in one of their api endpoints which actually resolves to Crunchyroll episode ids.

I assume that type="28" means that the resource "provider" is crunchyroll and the id in identifier tag matches the internal crunchyroll id of the (japanese) first episode of darling in the franxx.

I'm only concerned about the endpoints anidb provides to access the api. They publicly provide a SOAP or raw UDP endpoint which both are not ideal for modern development (imo).

[bytedream]

A AniDB integration would be indeed nice.

After a bit of searching multiple problems occurred. It is not guaranteed that the Crunchyroll title is the same as the title in AniDB. For example if you're located in Germany, only the german dub of sword art online is available and every season of it has a (DE) prefix or (German Dub) suffix (depending if your display language is set to german or english). The AniDB entries doesn't have this prefix/suffix.
Sure, crunchy-cli can try to remove any suspicious prefix/suffix, but I'm sure that this will lead to other problems where something gets removed which actually belongs to the correct title.
Another thing are special characters. In some cases titles have special characters like apostrophes. On Crunchyroll the title has ' as apostrophe, whereas AniDB has `. And here it's the same problem as described above, sure the character can be removed from the title but this may brake other things.

[bytedream]

Okay I have a question. For example, the first Re:Zero episode has E1A in it's title and the sequence number 0.5. With a option which handles such cases, should it return E1A or E0.5?

[BlastBolt5]

Well in this case I think it should just be dropped, Episode 1 here is a single episode that includes both episode 1a and 1b, But overall I think cause how letters are used by crunchyroll they should come out as be spit out as 1a and 1b

[bytedream]

I've added a option which should handle the most cases. You can test it by cloning the repo and checking out the feature/relative_sequence_number branch or using one of the pre-built binaries.
The new format option is called {relative_sequence_number} as it is based of the sequence_number api field but it might get renamed when I merge it, to state more clearly what it does.
It would be great if you guys could test this and give feedback @BlastBolt5 @alazare619 @ohitstom

[BlastBolt5]

Still doing some testing but this is what i'm seeing so far:

1. Forward slashes in in the formatting for -o do not appear for outputted file (can't sort into folders)
2. First episodes of a season with a .5 are formatted to E0 but later season .5 episodes overwrite the episode that should be downloaded, like E24.5 is downloaded as E24 and then the actual E24 just isn't downloaded
3. {relative_sequence_number} is not formatted to have 2 digits by default
4. Idk if this is related to this but when doing [S1E24-] for reincarnated as a slime episode 24.5 doesn't show up to be downloaded

Idk which parts of that is related to this specific branch but just thought i'd mention the issues

[ohitstom]

i think it would be nice if this option named anything that is a float as a special instead, that would solve 100% of the issue on my end since most of the time these episodes are classed as specials regardless, same goes for "1a" type episodes, if it returns sequence_number "anything.anything" make it a special

[BlastBolt5]

That also would probably be the best solution for me, and another option for 1a and 1b episodes would be outputting them as e1p1 and e1p2 possibly, but with how rare they are putting them in special is probably a good bet too

[bytedream]

Still doing some testing but this is what i'm seeing so far:

1. Forward slashes in in the formatting for -o do not appear for outputted file (can't sort into folders)

2. First episodes of a season with a .5 are formatted to E0 but later season .5 episodes overwrite the episode that should be downloaded, like E24.5 is downloaded as E24 and then the actual E24 just isn't downloaded

3. {relative_sequence_number} is not formatted to have 2 digits by default

4. Idk if this is related to this but when doing [S1E24-] for reincarnated as a slime episode 24.5 doesn't show up to be downloaded

Idk which parts of that is related to this specific branch but just thought i'd mention the issues

I think everything of the named issues are resolved now. Could you verify my changes?
(over at #257, that's the active PR)

[bytedream]

i think it would be nice if this option named anything that is a float as a special instead, that would solve 100% of the issue on my end since most of the time these episodes are classed as specials regardless, same goes for "1a" type episodes, if it returns sequence_number "anything.anything" make it a special

How should an episode be marked as special?

[BlastBolt5]

i think it would be nice if this option named anything that is a float as a special instead, that would solve 100% of the issue on my end since most of the time these episodes are classed as specials regardless, same goes for "1a" type episodes, if it returns sequence_number "anything.anything" make it a special

How should an episode be marked as special?

The best way would be if the season for that episode could be set to special or 00

[BlastBolt5]

Still doing some testing but this is what i'm seeing so far:

1. Forward slashes in in the formatting for -o do not appear for outputted file (can't sort into folders)
2. First episodes of a season with a .5 are formatted to E0 but later season .5 episodes overwrite the episode that should be downloaded, like E24.5 is downloaded as E24 and then the actual E24 just isn't downloaded
3. {relative_sequence_number} is not formatted to have 2 digits by default
4. Idk if this is related to this but when doing [S1E24-] for reincarnated as a slime episode 24.5 doesn't show up to be downloaded

Idk which parts of that is related to this specific branch but just thought i'd mention the issues

I think everything of the named issues are resolved now. Could you verify my changes? (over at #257, that's the active PR)

Yep just tested the new commit and all 4 of those issues are fixed, tysm!

[ohitstom]

i think it would be nice if this option named anything that is a float as a special instead, that would solve 100% of the issue on my end since most of the time these episodes are classed as specials regardless, same goes for "1a" type episodes, if it returns sequence_number "anything.anything" make it a special

How should an episode be marked as special?

if it is a float number, set its sequence number to just the word "special", since lots of metadata providers dont follow half episodes are more the common "special - title" format, such as adventure time (not applicable of course but just off the top of my head)

though this would likely fit better in its own flag that could work in tandem with relative sequence, maybe something named rename_floats or something

[bytedream]

Okay I see. I don't like the idea of an extra flag for this but I can't see another way atm.

To give a good amount of flexibility, the extra flag could be designed exactly like -o with the difference that it only gets used when the sequence number is a float.

E.g. (assuming that the new flag is called -os):

$ crunchy-cli archive -o "S{season_number}E{relative_sequence_number} - {title}.mkv" -os "Special - {title}.mkv" ...
# S1E1 - ep1.mkv
# S1E2 - ep2.mkv
# Special - ep2.5.mkv
# S1E3 - ep3.mkv

How do you thing about this?

[ohitstom]

yeah that looks great!

[BlastBolt5]

That looks awesome thank you so much

[BlastBolt5]

would there be anyway to make a flag to just ignore those episodes as well? Like how skip existing has one?

[bytedream]

I've pushed new code to the PR which adds the --special-output flag to control the output file name of special episodes as well as the --skip-specials flag to ignore special episodes entirely (what @BlastBolt5 suggested)

[bytedream]

Support for this landed in v3.1.0